Hypotheses about the relationship between a dependent variable (regressand, response, outcome) and a set of independent variables (regessors, predictors, features).
Model: \(\boldsymbol{y}=\boldsymbol{X\beta}+\boldsymbol{\epsilon}\)
Most commonly, \(\boldsymbol{\beta}\) is estimated with ordinary least squares (sum of squared residuals is minimized).
A typical questions are:
- How are variables A, B, C, … linked to variable Y?
- Can A, B, C, D … explain/predict Y?